本文選題：大數(shù)據(jù) + 流失預(yù)測��； 參考：《中國科學技術(shù)大學學報》2017年08期

【摘要】：電信客戶流失預(yù)測是電信運營商客戶關(guān)系管理系統(tǒng)的一個重要問題,其目的是預(yù)測具有較高流失風險的客戶.電信客戶流失預(yù)測模型的構(gòu)建過程包括數(shù)據(jù)預(yù)處理、不均衡處理、特征選擇和分類器的訓練與評估.針對電信數(shù)據(jù)集中存在的特征維度過高問題,結(jié)合過濾式特征選擇和嵌入式特征選擇方法的優(yōu)點,提出了一種基于Fisher比率和預(yù)測風險準則的分步特征提取方法.結(jié)合真實數(shù)據(jù)集的實驗結(jié)果表明,該方法能夠減少特征維度,提高分類器的預(yù)測效果.
[Abstract]:Customer churn prediction is an important problem in telecom operators' customer relationship management system, and its purpose is to predict customers with high risk of loss. The construction process of telecom customer churn prediction model includes data preprocessing, unbalanced processing, feature selection and classifier training and evaluation. Aiming at the problem of high feature dimension in telecom data set, a step feature extraction method based on Fisher ratio and predictive risk criterion is proposed, which combines the advantages of filtering feature selection and embedded feature selection. The experimental results of real data sets show that this method can reduce the feature dimension and improve the prediction effect of classifier.
【作者單位】： 中國科學技術(shù)大學自動化系;
【基金】：Supported by the National Natural Science Foundation of China(61375079)
【分類號】：F626
，

本文編號：1935618